Article Text

Download PDFPDF

Demand for health services and drug prescriptions among overweight or obese preschool children
  1. Maira Alejandra Ortiz-Pinto1,2,3,
  2. Honorato Ortiz-Marrón4,
  3. María D Esteban-Vasallo5,
  4. Agueda Quadrado-Mercadal6,
  5. Dayami Casanova-Pardomo7,
  6. Marta González-Alcón8,
  7. María Ordobás-Gavin4,
  8. Iñaki Galán1,2
  1. 1 Centro Nacional de Epidemiología, Instituto de Salud Carlos III, Madrid, Spain
  2. 2 Departmento de Medicina Preventiva y Salud Pública, Universidad Autonoma de Madrid/IdiPAZ, Madrid, Spain
  3. 3 Universidad del Norte, Barranquilla, Colombia
  4. 4 Servicio de Epidemiología, Consejería de Sanidad, Madrid, Spain
  5. 5 Servicio de Informes de Salud y Estudios, Consejería de Sanidad, Madrid, Spain
  6. 6 Centro de Salud Navas del Rey, Consejería de Sanidad, Madrid, Spain
  7. 7 Centro de Salud Cubas de la Sagra, Consejería de Sanidad, Madrid, Spain
  8. 8 Centro de Salud Lavapies, Consejería de Sanidad, Madrid, Spain
  1. Correspondence to Dr Iñaki Galán, Instituto de Salud Carlos III, Madrid 28029, Spain; igalan{at}


Objectives To evaluate the association between excess weight and the demand of health services in preschool children compared with healthy weight.

Methods The data come from the Longitudinal Study of Childhood Obesity cohort (1884 4-year-old children, residing in the Madrid region, Spain) who provided information through telephone questionnaire, physical examination and electronic medical records. We defined overweight, general and abdominal obesity based on body mass index, waist circumference and waist-to-height ratio. Using mixed models of multivariable negative binomial regression we calculated the incidence rate ratio (IRR) regarding primary care (PC) doctor visits, drug prescriptions and hospital admissions by weight status at the end of the 2-year follow-up.

Results Childhood general obesity was associated with a higher demand for PC services related to psychological problems (IRR=1.53; 95% CI 1.02 to 2.28) and childhood abdominal obesity, according to waist-to-height ratio, was related to more frequent problems of the musculoskeletal system (IRR=1.27; 95% CI 1.00 to 1.62). Drugs were prescribed more frequently to children falling under all three definitions of excess weight, compared with healthy weight children. No differences in the number of hospital admissions were observed.

Conclusions The demand of health services related to early childhood obesity was small. Nevertheless, obesity was associated with a slightly greater demand for drug prescriptions and for PC doctor visits related to psychological and musculoskeletal problems.

  • child obesity
  • waist circumference
  • body mass index
  • healthcare
  • drug prescriptions

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What is already known on this topic?

  • There is a perception that overweight and obesity have few clinical consequences in early life.

What this study adds?

  • Childhood obesity was associated with a slightly greater demand for psychological and musculoskeletal problems in primary care.

  • Childhood obesity was associated with more drug prescriptions.


In recent decades, the prevalence of obesity in children and adolescents has increased progressively worldwide, although since 2000 such a rise seems to have levelled off in many developed countries.1 By 2016, it corresponded to a disease burden of 33.6 and 27.3 Disability-Adjusted Life Years for 100 000 children aged 5–9 years and 10–14 years, respectively.2

Substantial evidence shows that excess weight in childhood and adolescence leads to adverse outcomes in adult life by increasing morbidity and the risk of premature mortality.3 Unfortunately, there is also evidence that some health professionals4 and parents5 underestimate the impact of childhood obesity, hampering the prevention and control of this health problem from an early age. This is due, in part, because childhood obesity, unlike its adult counterpart, does not substantially burden the health system regarding time and economic resources.6 7

There is scarce work evaluating the association between carrying some extra weight as a preschooler and the use of health services. For instance, based on parental self-reported comorbidity data, authors highlight the low burden of disease despite the high prevalence of excess weight.8 Others report an association between weight and health problems only among children with severe obesity.9 Still, Au10 concludes that children with excess weight generate higher pharmaceutical and medical costs at an early age than their healthy weight schoolmates.10

Based on literature reviews examining the effects of childhood obesity on the use of health services, authors report that the value of available evidence is limited by the heterogeneity in study designs, age groups, obesity definitions and outcome variables evaluated.11 Thus, it is necessary to develop more evidence on the clinical effects of childhood obesity, evaluating the demand and use of the health system prospectively by analysing longitudinal quality data.

The main objective of this study was to determine the prospective association between excess weight (overweight, general and abdominal obesity) and the use of specific health system services (visits to primary care (PC) physicians, drug prescriptions in PC and hospital admissions) among preschoolers. Our analyses are based on a representative sample of 4-year-old children and their healthcare utilisation during the 2 years of follow-up.

Materials and methods

Design and study population characteristics

The Longitudinal Study of Childhood Obesity, is a population-based cohort recruited in the Madrid region, an area of 6.5 million inhabitants in Spain. The objectives of the parent study are to: estimate weight changes between the ages of 4 and 14, describe the main risk factors associated with overweight and obesity, and evaluate the impact of excess weight on health status. The methodology and characteristics of the study sample have been previously reported.12 13 A total of 2627 children were included in the original cohort with parents acting as their proxy. Foreign parents and those with low educational level were less likely to fill out the baseline survey than others. For this study, our sample consisted of 1884 children examined and interviewed first at age 4 (baseline measurement) and then 2 years later, at age 6. Data were collected in three consecutive steps: first, a standardised physical examination performed at the health centre by one of the 31 participating paediatricians; second, a structured parental survey by computer-assisted telephone interview; and third, a review of electronic medical records to collect data on health service use during the 2-year follow-up.

Anthropometric measurements

During the physical examination weight, height and waist circumference were measured in a standardised manner by participating paediatricians. Weight was measured with a digital scale (SECA model 220, precision 0.1 kg), height was measured by telescopic height rod (SECA model 220, precision 1 mm) and waist circumference was measured with approved inextensible metric tape just above the iliac crests with the tape held horizontally and without tissue compression. Each variable was measured twice and the average of the two values was used for the analyses.

We calculated the child’s body mass index (BMI) (kg/m2), standardising its values according to the available WHO-2006 reference tables by age and sex. Based on the z-scores we classified children as overweight if their z-BMI>+1 SD and z-BMI ≤+2 SD) and as obese if their z-BMI score >+2 SD.14 The mean waist circumference readings were standardised based on age and sex reference tables for European–American children and adolescents proposed by Fernández et al15 interpolated monthly. Abdominal obesity was defined as any value ≥90th percentile according to the recommendations of the International Diabetes Federation.16 The waist-to-height ratio (waist circumference in cm/height in cm), was also used to define abdominal obesity when the value ≥90th percentile.17

Primary care electronic clinical history

Data on visits to doctor offices, drug prescription and recorded episodes of medical care (codified following the International Classification of Primary Care -2 classification) were collected from the Primary Care Electronic Clinical history. We analysed the information corresponding to the most commonly reported events, that is, those related to the respiratory, musculoskeletal and nervous systems as well as data on psychological problems. The records of hospital discharges were obtained through the Minimum Basic Data Set, a single file of hospital admissions/discharges from all public and private hospitals.


The sociodemographic covariates included in the models were sex, child’s age in months, mother’s educational achievement (five categories), duration of breastfeeding (four categories), family purchasing power (low (0–5 points), medium (6–7 points) and high (8–9 points)) defined based on the Family Affluence Scale18 and parent’s perception of the child’s health status at baseline classified as either optimal (very good or good) or suboptimal health (regular, poor, or very poor).

Statistical analyses

The initial sample consisted of 1884 participants. We excluded 27 children classified as underweight and/or with no waist circumference data at baseline. After taking into account missing values for some other covariates, the final sample was 1857 for analyses regarding BMI-based excess weight or 1851 participants for analyses related to abdominal obesity.

First, we performed descriptive analyses and simple statistical hypothesis testing at the bivariate level (Student's t-test or analysis of variance depending on the number of groups to contrast). The null hypothesis assumed no difference in the average (and variance) number of PC visits, of drug prescriptions and of hospital admissions for each variable and corresponding categories. Second, since over-dispersion was detected, we designed multilevel models of negative binomial regression, including the identification of the paediatrician as a random factor, to evaluate the number of PC visits, drugs prescriptions and hospital admissions during the 2-year follow-up period. The quotient of incidence rates (IRR) and their respective 95% confidence intervals and p values were calculated. All models were adjusted for the covariates described above. We tested for interactions between our excess weight variables and sex as well as family purchasing power.

The level of statistical significance was set at a p value of <0.05. All the analyses were performed using Stata V.14. (StataCorp. 2015. Stata Statistical Software: Release 14).


The characteristics of the sample are described in table 1. From age 4 to age six our participants had an accumulated average of 11.7 visits to a PC physician, 1.25 drug prescriptions from PC physicians and 0.22 hospitalisations. We observed some statistically significant differences; first, a greater number of drug prescriptions and hospital admissions among boys than girls and, second, a lower frequency of PC doctor visits among children of college-educated mothers and high socioeconomic status than other children. As one might expect, children with a suboptimal health status made greater use of health services.

Table 1

Sample characteristics according to type of healthcare use: primary care visits, drug prescriptions and hospital admissions for between baseline (age 4) and follow-up (age 6)

Table 2 shows the association between excess weight and PC doctors’ visits, drug prescriptions and hospital admissions. Children affected by general or abdominal obesity at age 4 visited a PC doctor a statistically similar number of times as normo-weight children (IRR=1.08 and 1.06, respectively). However, children falling under any of the three excess weight categories were prescribed a greater number of drugs than their normo-weight counterparts (IRR=1.62; 95% CI 1.11 to 2.38) for general obesity, 1.37 (95%CI 0.99 to 1.90) for abdominal obesity according to waist circumference (not statistically significant) and 1.34 (95%CI 1.02 to 1.76) for abdominal obesity according to the waist-to-height ratio. Finally, we observed no differences by baseline weight status in the incidence of hospital admissions during the follow-up period.

Table 2

Weight status and abdominal obesity at age 4 according to type of healthcare use: primary care visits, drug prescriptions and hospital admissions, during the 2-year follow-up

The average number of PC visits by body system during the 2-year follow-up was: four visits for respiratory system problems, 0.5 for musculoskeletal system problems, 0.2 for nervous system problems and 0.5 for psychological problems (data not shown). Table 3 describes the relationship between weight status and PC visits grouped by different body systems. Results showed an increase in the cumulative incidence rate for musculoskeletal system problems (IRR=1.27; 95% CI 1.00 to 1.62) associated with abdominal obesity based on waist-to-height ratio but not with the other two obesity indicators. We also observed a positive relationship between childhood BMI-based general obesity and PC visits due to psychological problems (IRR=1.53; 95% CI 1.02 to 2.28). However, the association with other obesity indicators failed to reach statistical significance. We did not find associations with respiratory or nervous systems problems.

Table 3

Association between weight status and abdominal obesity at age 4 and primary care visits during the 2-year follow-up. Primary care visits are classified according to the four systems most health problems relate to

None of the interactions explored reached statistical significance.


Our study shows that early childhood obesity (both general and abdominal) is associated with a higher demand for drug prescriptions during a 2 year follow-up period. Although we failed to detect a higher overall demand for PC services associated with obesity, additional doctor’s visits were probably related to psychological and musculoskeletal problems. No differences were observed regarding the number of hospital admissions.

Childhood obesity may adversely affect practically every system in the organism19 to the point that large literature reviews link childhood and adolescent obesity to reduced self-rated health and quality of life.20 21 However, there is a discrepancy between said reports and the low burden of disease detected by parents.5 Additionally, there is evidence that some health professionals underestimate the impact of childhood obesity,4 hampering any excess weight prevention and control efforts at early ages.

Previous studies show an increase in demand for health services among children and adolescents affected by obesity. The magnitude of such increases tend to be small-to-moderate and mostly reflect demand for services other than hospital admissions.7 22–26 Estabrooks and colleagues observed an 11% increase in the risk of PC visits during a 1-year follow-up and of 6% during a 3 year follow-up of a sample of 3-year-olds to 17-year-olds.22 Although we detected an 8% increase in visits, but if failed to reach statistical significance. Janicke and co-authors26 observed a slightly higher increase (19%) in children between 7 and 15 years of age; whereas Hampl et al,27 failed to detect any significant increase in PC visits in the 5-year-old to 18-year-old population.

Obesity, based on waist circumference/height, was associated with a greater demand for PC services involving musculoskeletal system issues. Childhood obesity’s impact on a young musculoskeletal system works through the dysfunction of the joints. This in turn, generates foot, ankle and knee problems, as well as muscle pain and increased risk of fractures,28 though the evidence regarding the risk of fractures and accidents remains scarce.29 The only longitudinal study of 2-year-old preschool children revealed an increase in hospitalisations due to musculoskeletal problems during the 3-year follow-up among children affected by obesity.24 The evidence regarding a link between obesity and risk of accidents remains inconclusive. Whereas Lynch et al 23 reported that obesity seemed to increase such risk, Ferro and colleagues30 concluded that obesity played a protective role. Such role, however, probably reflects the inverse relationship between obesity and physical activity levels among children.

The finding that BMI-based obesity was also associated with greater demand for PC services related to psychological problems supports previous findings identifying poor mental health as one of the most widespread side-effects of childhood obesity.4 However, few studies evaluate obesity’s impact on psychological-related health services use. Estabrook and Shetterly22 and Turer and colleagues25 observed an increase in visits due to mental health problems in children aged 3–17 and 10–17 years, respectively, with excess weight. Wooldford et al 31 found an increased hospital burden due to affective disorders among the 2-year-old to 18-year-old patient population with a secondary hospital diagnosis of obesity.

The greater number of drugs prescribed to children affected by obesity in our study also supports previous findings such as Solmi and Morris’s32 on prescription use in a cohort of 5-year-olds followed for 7 years. In another study10 conducted on Australian 4-year-old to 5-year-old children, excess weight was associated with higher pharmaceutical expenditure. However, Hayes et al 24 analysed health records data from 2-year-olds to 5-year-olds and failed to detect differences in medication use by weight category, although the small sample size may be the culprit.

Again, published evidence regarding the relationship between obesity and increased risk of hospital admissions is inconclusive. Three studies found no association,9 10 23 whereas two others reported a direct relationship between risk of admissions and excess weight in children.6 24 One of these works6 linked BMI trajectories to health services use and concluded that children who gained weight quickly, that is, reaching high BMI values quickly and maintaining that excess weight for up to 10 years, were more likely to be hospitalised during the study period.

Finally, we explored whether changes in weight status during the 2-year follow-up may have influenced our results. We repeated the analysis only with children classified under the same weight category at age 4 and again at age 6 (data not shown). The association between obesity and greater number of prescriptions or psychological problems with obesity remained unchanged. However, the association between musculoskeletal problems and excess weight was no longer significant.

Limitations and strengths of the study

When interpreting our study’s results, some limitations should be kept in mind. First, foreign parents or those with low educational achievement were less likely to fill out the baseline survey. This moderate selection bias may limit the generalisation of results to the entire population of 4-year-olds in the Madrid region. Second, the short follow-up time (2 years), the small sample size and the limited number of events reduced the statistical power of some estimates.

Third, there are no validation studies of the clinical data included in the electronic medical record for the region’s children, although errors are probably random, that is, affecting obese and non-obese children equally. Finally, incompatibility between data systems precluded extending the analysis to demand for emergency room services.

Our study has important strengths. First, its longitudinal design, and adjustment for key likely confounders, offer great potential for the detection and establishment of causal associations. Second, it contributes to the very scarce body of work examining the use of health services by preschool children. Third, the anthropometric measures were performed in an objective and standardised manner. Further, general and abdominal obesity were differentiated for the first time when examining their relation to the use of health services. Finally, the models were adjusted for the main sociodemographic covariates, as well as for duration of breastfeeding and the child’s health status as perceived by the parents.

In sum, the demand for health services related to obesity was small in preschool age. Nevertheless, general and abdominal obesity were associated with a slightly greater burden on health systems due to an increased risk of PC visits related to musculoskeletal and psychological problems, as well as to a higher number of drug prescriptions in primary care.


The authors thank all participating families, the paediatricians from the Network of Sentinel Doctors of the Community of Madrid and the two companies interviewing the families: Demométrica SL and Sondaxe SL.



  • Contributors MAO-P and IG, conceptualised and designed the study, drafted the initial manuscript, and reviewed and revised the manuscript. MAO-P and IG, designed the data collection instruments, collected data, carried out the initial analyses and MAO-P, HO-M, ME-V and IG, reviewed and revised the manuscript. HO-M, ME-V, AQ-M, DC-P, MG-A, MO-G, collected data, and critically reviewed the manuscript for important intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval The study protocol was approved by the Ethics Committee of the Ramón y Cajal Hospital in Madrid, approval number 242 and register number 44/10.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement There are no data in this work. Data are available upon reasonable request. Data may be obtained from a third party and are not publicly available. No data are available. All data relevant to the study are included in the article or uploaded as supplementary information.